High-Spin BBH Subpopulation from AGN Accretion

Technical Summary: A High-Spin Subpopulation in Binary Black Hole Mergers Consistent with AGN Accretion

Problem Statement
The formation environments of merging binary black holes (BBHs) remain uncertain. While standard stellar evolution predicts an upper mass limit of $\sim 50 M_\odot$ due to pair-instability supernovae, gravitational-wave observations have revealed mergers with component masses exceeding this limit. Two primary mechanisms have been proposed to explain these massive systems: hierarchical mergers in dense stellar environments (e.g., globular clusters, nuclear star clusters) and mass growth via gas accretion in Active Galactic Nucleus (AGN) disks. While hierarchical mergers are widely explored, the accretion channel has received less observational scrutiny. Furthermore, existing population analyses have not systematically assessed the role of accretion-driven growth within a unified statistical framework that competes directly with hierarchical models. A key challenge is distinguishing between these channels, as both can produce high-mass systems, but they predict distinct spin signatures: hierarchical mergers typically yield remnant spins near $a \approx 0.7$, whereas sustained accretion can spin black holes up to $a \approx 0.9$.

Methodology
The authors employ a minimalist, population-level approach to search for an accretion-origin subpopulation using only spin magnitudes ($a_1, a_2$) from 166 binary black hole mergers in the GWTC-2.1, GWTC-3, and GWTC-4.0 catalogs.

• Mixture Model: They construct a three-component mixture model representing three formation channels:
  1. Standard: Low natal spins (isolated binary evolution, dynamical capture), modeled as a Beta distribution peaked at zero.
  2. Hierarchical: Second-generation–first-generation (2g–1g) mergers, modeled with a primary spin distribution centered at $a_1 \approx 0.7$ (based on numerical relativity predictions) and a low-spin secondary.
  3. Accretion: Black holes spun up by sustained gas accretion in AGN disks, modeled with both spins following a truncated normal distribution centered near the theoretical equilibrium of $a \approx 0.9$.
• Inference Strategy: The authors fix the shapes of the component distributions based on theoretical predictions and infer only the mixing fractions ($f_{\text{hier}}$ and $f_{\text{acc}}$) and the low-spin shape parameter ($\beta_{\text{low}}$) using hierarchical Bayesian inference. This approach avoids assumptions about mass distributions, which are currently poorly constrained for these channels.
• Data Handling: The analysis utilizes posterior samples from the NRSur7dq4 numerical relativity surrogate model where available (90 events) and mixed-approximant samples for the remainder. Selection effects are corrected using importance-sampling reweighting against the LVK O3+O4a injection set.
• Validation: Post hoc validation is performed by examining properties (mass, effective spin $\chi_{\text{eff}}$, mass ratio) of the inferred accretion candidates, which were not used in the fitting process.

Key Results

• Evidence for Accretion: The analysis yields strong evidence ($\ln \mathcal{B}_{M1/M0} = 5.7$) for a subpopulation of accretion-origin mergers. The inferred fraction of accretion candidates is $\sim 10\%$ (90% credible interval: $1\% - 14\%$), with a maximum a posteriori (MAP) estimate of $8.2\%$.
• Spin Distribution: The accretion component is characterized by primary spins clustered near $a_1 \approx 0.9$, well-separated from the bulk population.
• Ruling Out Hierarchical Origin: The hierarchical-merger prediction of $a_1 \approx 0.7$ is decisively disfavored as the location of the high-spin subpopulation ($\ln \mathcal{B}_{M1/M2} = 5.7$). The data do not support a significant detection of the hierarchical channel from spin magnitudes alone.
• Candidate Properties: Post hoc analysis reveals that accretion candidates ($p_{\text{acc}} > 0.3$) have systematically higher primary masses (median $m_1 = 58 M_\odot$) and higher effective aligned spins (median $\chi_{\text{eff}} = 0.33$) compared to the standard-dominated population.
• Specific Events:
  • GW231123: The most massive BBH merger detected to date is the most confident accretion candidate ($p_{\text{acc}} > 0.97$), with high spins ($a_1 \approx 0.90, a_2 \approx 0.91$).
  • GW190517: A candidate with $m_1 \approx 39 M_\odot$ (below the pair-instability gap) demonstrates that accretion spin-up operates across a range of masses, not just in the high-mass regime.
  • GW190521: Previously interpreted as a hierarchical merger, this event shows comparable support for an accretion origin ($p_{\text{acc}} = 0.54$), with broad spin posteriors supporting high values.

Robustness
The results are robust to variations in the assumed width of the accretion peak, the functional form of the standard component (tested with flexible Beta distributions), and the choice of waveform approximants. The detection of the accretion subpopulation persists across all tested configurations.

Significance and Claims
The paper claims to provide the first population-level observational evidence for an accretion-origin subpopulation in binary black hole mergers. By isolating spin magnitudes and fixing component shapes to theoretical predictions, the authors demonstrate that the observed high-spin tail in the GWTC-4 population is consistent with growth via gas accretion in AGN disks. The inferred fraction of $\sim 10\%$ aligns with theoretical estimates of the AGN contribution to BBH merger rates. The study suggests that the high-spin, high-mass population is not solely the product of hierarchical mergers but includes a significant component driven by accretion, which naturally links high spins with near-equal mass ratios and aligned spins. The authors note that while their spin-only analysis cannot significantly detect the hierarchical fraction, their results complement existing findings on $q$–$\chi_{\text{eff}}$ correlations and provide a physical decomposition of the spin structure observed in recent catalogs.